Early identification of total hip arthroplasty devices with increased risk of failure can be challenging because devices consist of multiple components and hundreds of distinct components are currently used in surgery. Ideally, a method can identify individual components with an increased risk of revision surgery using a time-to-event endpoint while also limiting the confounding effects of other components in the device and patient characteristics. In this talk two machine learning methods are considered for this problem, regularized/unregularized Cox models and random survival forest. The approaches are illustrated using data are from 74,520 implantations and 348 unique components used among elective primary total hip replacements in the Kaiser Permanente Total Joint Replacement Registry. The results and practical considerations favor the use of regularized/unregularized Cox models. Future applications are discussed.